Colchicine for the Secondary Prevention of Cardiovascular Diseases: A Cumulative-Dose Meta-analysis of Randomized Controlled Trials including 31,397 Subjects Worldwide

Abstract

Background

Colchicine has been incorporated into major clinical guidelines for the secondary prevention of cardiovascular disease (CVD). However, recent randomized trials have presented contradictory results.

Objective

We aimed to synthesize the current evidence on colchicine in secondary CVD protection, using a cumulative-dose approach.

Methods

We conducted a meta-analysis incorporating all randomized controlled trials (RCTs) globally. RCTs directly comparing colchicine versus placebo/standard care for the secondary prevention of cerebrovascular or coronary vascular disease were included. Odds ratios (OR) were derived for the primary outcome, defined as the prospective occurrence of major adverse cardiovascular events (MACE). Secondary outcomes included mortality, individual components of MACE, C-reactive protein, and adverse effects.

Results

In total, 14 RCTs including 31,397 participants were included. Colchicine significantly reduced MACE (OR 0.80; 95% confidence interval [CI] 0.68–0.94) in both acute atherothrombotic CVD and all CVD (OR 0.72; 95% CI 0.60–0.86) and resulted in significant prospective reductions in C-reactive protein. The threshold effect was apparent, with a protective benefit of colchicine against MACE at higher cumulative exposure ≥ 90 mg-days (OR 0.66; 95% CI 0.52–0.84). Colchicine resulted in no differences in cardiovascular or non-cardiovascular mortality.

Conclusions

Colchicine significantly reduces MACE in both acute atherothrombotic and all CVD across multiple ethnicities, with a threshold protective effect that clinically corresponds to treatment with 0.5 mg daily for at least 6 months. Importantly, there was no signal of increased all-cause mortality.

Registration

PROSPERO identifier no. CRD420251003142.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40256-025-00743-y.

Key Points

In this meta-analysis of randomized controlled trials (RCTs) including 31,397 participants globally, colchicine significantly reduced the incidence of major adverse cardiovascular events in both acute atherothrombotic and all cardiovascular disease across multiple ethnicities.

Cumulative-dose analyses showed a threshold protective effect of colchicine, clinically corresponding to treatment with 0.5 mg daily for at least 6 months. There was no signal of increased all-cause or non-cardiovascular mortality.

This study provides novel data that substantiate current clinical guideline recommendations.

Introduction

Despite clinical advances, cardiovascular disease (CVD) remains the leading cause of death globally [1], and the need for control cannot be overstated. Beyond conventional risk factors, systemic inflammation potentiates the development of atherosclerosis [2, 3], which manifests clinically as ischemia in the heart, cerebral circulation, and other peripheral end organs. Acute atherothrombotic events, pathophysiologically characterized by plaque rupture, are often triggered by inflammation, which is promulgated in a vicious cycle. Clinical control of inflammation with various therapeutic agents has been shown to reduce both the incidence [4] and the recurrence [5] of CVD events by ameliorating residual inflammation. However, these individual therapies have their limitations, including clinical and cost effectiveness, patient affordability, and acceptability.

Colchicine, originally derived from the plant Colchicum autumnale, has been used traditionally for medicinal purposes for millennia. It is now clinically used mostly for the treatment of gout, pericarditis, and familial Mediterranean fever, at a low cost [6]. It is thought to exert pharmacological benefits by disrupting microtubule polymerization [7] and inhibiting neutrophil activation and NLRP3 inflammasome, thereby lowering key cytokines such as interleukin (IL)-1β and IL-6. These provide its mechanistic basis to help stabilize plaques by reducing inflammation and protecting endothelial cells [7]. Randomized studies have suggested that colchicine significantly reduces the risk of cardiovascular events in patients with CVD. Indeed, it has been incorporated into major clinical guidelines for secondary protection against CVD. Nevertheless, recent large clinical trials reported conflicting results, including negative trials for large samples of patients with acute myocardial infarction (MI) [8] or ischemic stroke [9, 10]. Moreover, the dosage and treatment durations of studies of colchicine have been variable. Furthermore, concerns have been raised regarding potentially increased rates of all-cause, particularly non-cardiovascular, mortality in patients treated with colchicine [6, 11–14]. As such, there is a need to clarify the net clinical benefits of colchicine in the secondary prevention of CVD.

Therefore, we carried out a meta-analysis incorporating all available evidence from randomized controlled trials (RCTs) and adopted a cumulative-dose approach to derive an overall interpretation of the net clinical benefits versus harms of colchicine in the secondary prevention of CVD across a range of ethnicities worldwide.

Methods

Search Strategy and Study Selection

We performed a systematic review and meta-analysis according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15] to investigate the effect of colchicine on the risk of major adverse cardiovascular events (MACE) in patients with established CVD. The spectrum included acute atherothrombotic CVD and all CVD. Acute atherothrombotic CVD included all acute ischemic vascular events of the coronary or cerebrovascular circulations, that entailed acute coronary syndrome (ACS), ST-segment elevation or non-ST-segment elevation MI, ischemic stroke, and transient ischemic attack (TIA). All CVD included any form of acute atherothrombotic CVD and stable coronary artery disease (CAD). The following keywords were used in the search: colchicine, stroke, major adverse cardiac events (MACE), myocardial infarction, ACS, and cardiac. No restrictions were placed on publication year, and studies published in English, French, German, Italian, Spanish, and Chinese were included. Eligible studies that involved the above-defined atherothrombotic CVD or all CVD were included. Studies that involved patients primarily with atrial fibrillation, pericarditis, or other cardiac conditions were excluded. The literature search was performed in the PubMed, Scopus, Embase, and Cochrane Library databases, with the last search conducted on February 1, 2025. All searches were performed twice by two independent investigators (HYL and TKC) and further verified by a third senior investigator (YHC). All included studies were reviewed independently and assessed for quality. Studies had to compare colchicine with a placebo or standard care, have a follow-up period of at least 1 month, and report incident occurrences of MACE as an outcome. Only fully published manuscripts on RCTs were included in the analyses. Conference abstracts, case reports, reviews, and observational studies were excluded. The CONSORT checklist was used to assess the quality of RCTs [16]. The Cochrane risk-of-bias assessment tool (RoB2) [17] was used to evaluate the risk of bias. Studies with a high risk of bias were excluded from the analysis to avoid the inclusion of poor-quality studies.

Ethics Approval

As this was a meta-analysis of published studies, there was no direct patient clinical contact. Therefore, no specific ethical approval was obtained or deemed necessary. This meta-analysis was registered with the PROSPERO International prospective register of systematic reviews (CRD420251003142).

Definition of Outcomes

The primary outcome was defined as any incident occurrence of MACE in patients with acute atherothrombotic CVD and those within the spectrum of all CVD. Secondary outcomes included all-cause or cardiovascular/non-cardiovascular, individual components of MACE, prospective changes in C-reactive protein (CRP), and adverse effects. As the definitions of MACE varied across the included studies, we standardized our primary outcome definition of MACE to include the following components, where available, to ensure consistency: ischemic stroke, ACS, MI, unstable angina, coronary revascularization, and cardiovascular death. Each individual MACE component was extracted as defined by the original trial authors from the full-text publications and supplementary materials. Adverse effects included infections, deranged liver/renal function, neuropathy, myalgia, myelosuppression, and gastrointestinal (GI) adverse effects. We extracted the following elements: raw numbers of participants, primary outcomes, and secondary outcomes; study characteristics, such as trial acronyms, year of publication, study design, and follow-up period; and baseline data, such as primary disease, sex distribution, mean age, and comorbidities, including hypertension, diabetes mellitus, and smoking status.

Definition of Cumulative Colchicine Exposure

We extracted data from RCTs on colchicine dosage and treatment duration, then quantified cumulative colchicine exposure based on a dosage–duration composite estimate: [colchicine in mg/day × number of days of intervention]. The median value of cumulative exposure was 90 mg-days, which was a priori defined as the cut-off for comparison.

Changes in CRP

To aid mechanistic understanding, we performed a meta-analysis of highly sensitive CRP (hsCRP)/CRP levels, which calculated mean differences between the colchicine and control groups to derive treatment effects of colchicine on pre-treatment, post-treatment, and prospective changes in hsCRP/CRP levels.

Statistical Analysis

We used the “meta” package in R Studio version 2024 for all analyses. The random-effect Mantel-Haenszel method was used to derive odds ratio (ORs) with 95% confidence interval (CIs) estimates of the primary and secondary outcomes, and the DerSimonian–Laird method was used to estimate the between-studies variance. To address zero events in outcomes, the Mantel-Haenszel method applied a small continuity correction by adding 0.5 to avoid computational issues with zero events in one arm. Studies with zero events in both arms were excluded from the calculation of pooled ORs [18]. Heterogeneity was assessed using the I² statistic and Cochran’s Q test. Low heterogeneity was defined as I² = 0–40% with a Q test p-value ≥ 0.10, moderate heterogeneity as I² = 30–60%, and substantial to considerable heterogeneity as I² = 50–100% [18]. Forest plots were used to visualize the effects of each outcome.

We conducted prespecified subgroup analyses using the Chi-squared test for group differences to assess the effects of study and clinical characteristics. Meta-regression analysis was performed for outcomes with more than 10 studies with moderate to high heterogeneity to identify potential sources of variability and explore the influence of study-level characteristics on the results. Sensitivity analyses were conducted using the leave-one-out method to assess the influence of each study and by excluding studies rated as with “some concerns” in the RoB2 assessment to evaluate the impact of study quality on the primary findings. Publication bias was assessed using funnel plots and Egger’s test. For comparison, studies that reported only medians and interquartile ranges were transformed into means and standard deviations according to Abbas et al. [19].

Results

The search of the medical databases resulted in 728 records (Fig. 1 in the electronic supplementary material [ESM]). After removing 201 duplicates, we undertook title and abstract screening of 527 studies and removed 508 studies because of prima facie evidence of being out of scope, being not relevant, or having incomplete data. Following the full-text screening of the remaining 19 studies, four were excluded because they were sub-studies of the included RCTs, and one study was excluded because it did not report MACE recurrence. Finally, 14 RCTs published between 2012 and 2024, with a total of 31,397 participants, were included in the meta-analysis [8–10, 20–31]. According to the RoB2, 11 studies had a low risk of bias, and three were categorized as having “some concerns” because the randomization procedures were unclear (Fig. 2 in the ESM). A summary of the included studies is presented in Table 1. Baseline characteristics of each study are summarized in Table 2.

Fig. 1.

Effect of colchicine on the primary outcome of MACE. Colchicine reduces MACE in patients with acute atherothrombotic CVD (OR 0.80; 95% CI 0.68–0.94; heterogeneity I² = 42.8%; p = 0.06) and patients with all CVD (OR 0.72; 95% CI 0.60–0.86; I² = 62.1%, p = 0.001). CI confidence interval, CVD cardiovascular disease, MACE major adverse cardiovascular events, OR odds ratio

Table 1.

Summary of included studies

Study, year; country	Trial acronym; study design	Study population	Treatment/control comparisons	COL dosage	FU period (mo.)	Sample size (treatment/control)
Jolly et al. [8], 2024; 14 countries	CLEAR; DB RCT	MI	COL/PL	0.5 mg BID for pts >70 kg and OD for <70 kg for 90 d; then, 0.5 mg OD for all	36	3528/3534
Kelly et al. [9], 2024; 13 European countries, Canada	CONVINCE; OL RCT	IS/TIA	COL/SC	0.5 mg OD	33.6	1569/1575
Li et al. [10], 2024; China	CHANCE-3; DB RCT	IS/TIA	COL/PL	0.5 mg BID for 3 d, then 0.5 mg OD	3	4176/4167
Khiali et al. [20], 2024; Iran	NA; DB RCT	STEMI	COL + EMP/EMP only	0.5 mg BID	3	35/36
Bouleti et al. [21], 2024; France	CONVERT-MI [22]; DB RCT	STEMI	COL/PL	2 mg LD, then 0.5 mg BID for 5 d	12	101/91
Akrami et al. [23], 2021; Iran	NA; DB RCT	ACS	COL/PL	0.5 mg OD	6	120/129
Nidorf et al. [24], 2020; Australia, Netherlands	LoDoCo2; DB RCT	CAD	COL/PL	0.5 mg OD	28.6	2762/2760
Tong et al. [25], 2020; Australia	COPS; DB RCT	ACS	COL/PL	0.5 mg BID for 1 mo, then 0.5 mg OD	12	396/399
Shah et al. [26], 2020; USA	COL-PCI; DB RCT	ACS/CAD	COL/PL	1.8 mg single dose before PCI	1	206/194
Hennessy et al. [27], 2019; Australia	LoDoCo-MI; DB RCT	MI	COL/PL	0.5 mg OD	1	111/113
Tardif et al. [28], 2019; 12 countries	COLCOT; DB RCT	MI	COL/PL	0.5 mg OD	22.6	2366/2379
Akodad et al. [29], 2017; France	COLIN; OL RCT	Acute STEMI underwent primary PCI	COL/SC	1 mg OD	1	23/21
Nidorf et al. [30], 2013; Australia	LoDoCo; OL RCT	CAD	COL/SC	0.5 mg OD	36	282/250
Raju et al. [31], 2012; Australia	COOL; DB RCT	ACS/IS	COL/PL	1 mg OD	1	36/38

ACS acute coronary syndrome, BID twice daily, CAD coronary artery disease, COL colchicine, d day(s), DB double blind, EMP empagliflozin, FU follow-up, IS ischemic stroke, LD loading dose, MI myocardial infarction, mo month, NA not applicable, OD once daily, OL open label, PCI percutaneous coronary intervention, PL placebo, PTS patients, RCT randomized controlled trial, SC standard care, STEMI ST-elevated MI, TIA transient ischemic attack

Table 2.

Baseline characteristics of included studies

Study	Trial acronym	Age, y	Male	HT	DM	Smoking
Jolly et al. [8], 2024	CLEAR	61 ± 10	80	46	18	41
Kelly et al. [9], 2024	CONVINCE	66 ±10	70	65	22	22
Li et al. [10], 2024	CHANCE-3	69 ±11	62	77	32	24
Khiali et al. [20], 2024	N.A.	60 ±10	73	55	0	39
Bouleti et al. [21], 2024	CONVERT-MI	60 ±10	80	31	13	43
Akrami et al. [23], 2021	N.A.	57 ± 8	69	45	24	41
Nidorf et al. [24], 2020	LoDoCo2	66 ± 9	85	51	23	12
Tong et al. [25], 2020	COPS	60 ±10	79	50	19	35
Shah et al. [26], 2020	COLCHICINE-PCI	66 ± 10	94	92	58	22
Hennessy et al. [27], 2019	LoDoCo-MI	61 ± 13	77	47	22	60
Tardif et al. [28], 2019	COLCOT	61 ± 11	81	51	20	30
Akodad et al. [29], 2017	COLIN	60 ± 12	80	43	14	70
Nidorf et al. [30], 2013	LoDoCo	66 ± 9	89	NA	30	5
Raju et al. [31], 2012	COOL	57 ±10	89	43	16	44

Data are presented as mean ± standard deviation or % unless otherwise indicated.

DM diabetes mellitus, HT hypertension, MACE major adverse cardiovascular events, NA not applicable, y years

Colchicine Reduces MACE in Acute Atherothrombotic CVD and All CVD Populations

The pooled analysis demonstrated that treatment with colchicine significantly reduced MACE compared with control (Fig. 1). Such results remained consistent in both the acute atherothrombotic CVD population and in all CVD populations. For patients with acute atherothrombotic CVD, 12 trials involving 12,564 patients receiving colchicine and 12,577 controls were studied. These resulted in a pooled OR of 0.80 (95% CI 0.68–0.94) for MACE among patients treated with colchicine. The heterogeneity among these studies was moderate (I² = 42.8%, p = 0.06). In the broader populations including all CVD, 14 trials comprising 15,711 patients in the colchicine group and 15,686 in the control group were included. The pooled OR for MACE in patients treated with colchicine was 0.72 (95% CI 0.60–0.86). This analysis showed a higher degree of heterogeneity (I² = 62.1%, p = 0.001). Overall, the findings indicated the benefit of colchicine for secondary prevention against MACE was consistent between acute atherothrombotic and all CVD populations.

Cumulative Dosing of Colchicine and Secondary CVD Protection

Prespecified analyses were performed for the primary outcome by categorizing cumulative colchicine exposure according to the median value of ≥ 90 mg-days (Fig. 2). All patients with CVD who received cumulative colchicine exposure of 90 mg-days exhibited a significant reduction in MACE (OR 0.66; 95% CI 0.52–0.84). Heterogeneity was considerable (I² = 76%). In comparison, patients who received cumulative colchicine exposure < 90 mg-days did not demonstrate significant benefits in terms of MACE prevention (OR 0.93; 95% CI 0.79–1.09), with no significant heterogeneity. The between-groups difference was statistically significant (p = 0.02).

Fig. 2.

Threshold effect of cumulative colchicine exposure on the primary outcome. A threshold effect of cumulative exposure was observed on the secondary protection of colchicine against MACE (cumulative colchicine exposure ≥ 90 mg-days: OR 0.66; 95% CI 0.52–0.84 vs <90 mg-days: OR 0.93; 95% CI 0.79–1.09, between-groups difference p = 0.02). CI confidence interval, MACE major adverse cardiovascular events, OR odds ratio

Studies were also stratified according to the respective duration and dosage of colchicine used (Fig. 3 in the ESM). For studies with treatment durations of < 1 week or of 1–3 months, the pooled ORs were, respectively, 0.69 (95% CI 0.39–1.21) and 0.95 (95% CI 0.80–1.13), indicating no statistically significant differences (no heterogeneity, I² = 0%). However, MACE was significantly reduced in patients receiving colchicine for ≥ 6 months (OR 0.65; 95% CI 0.51–0.83; significant heterogeneity, I² = 79%). The between-groups difference was statistically significant (p = 0.04).

For subgroup analysis according to daily dose alone (Fig. 4 in the ESM), 9 of 14 studies had a daily dosage of colchicine of 0.5 mg. This included the vast majority of the sample size in this meta-analysis. The COLCHICINE-PCI trial [26] was ruled out because of its single-dose nature. Analyses revealed that daily doses of 0.5 mg were associated with a significant reduction in MACE (OR 0.70; 95% CI 0.57–0.86), though heterogeneity was considerable (I² = 70%), whereas a higher daily dose of 1 mg accounted for a very small sample (n = 381) and did not show a significant difference (OR 0.75; 95% CI 0.45–1.27).

Secondary Outcomes: Individual Components of MACE

Colchicine significantly reduced the prospective occurrence of several individual components of MACE (Fig. 3). In six studies in ACS, colchicine significantly lowered the risk of ACS (OR 0.49; 95% CI 0.27–0.89). Similarly, among 10 studies in MI, colchicine significantly reduced the prospective risk of MI (OR 0.77; 95% CI 0.62–0.95). Five studies assessed unstable angina, where colchicine also significantly reduced the risk of unstable angina (OR 0.39; 95% CI 0.19–0.77). Additionally, pooling five studies with relevant data, colchicine also significantly reduced the incident risk of coronary revascularization (OR 0.70; 95% CI 0.49–0.99). Ischemic stroke was reported in nine studies. Colchicine appeared to be associated with a lower risk of ischemic stroke, but statistical significance was borderline (OR 0.82; 95% CI 0.64–1.05). Heterogeneity across these outcomes ranged from low to moderate (I² = 26–72%), with higher variability in ACS and coronary revascularization. Overall, these findings suggested that colchicine effectively reduced the risk of overall and most individual components of MACE, with generally consistent results across studies.

Fig. 3.

Effects of colchicine on individual components of MACE. Colchicine significantly lowered the risk of ACS (OR 0.49; 95% CI 0.27–0.89), MI (OR 0.77; 95% CI 0.62–0.95), unstable angina (OR 0.39; 95% CI 0.19–0.77), and coronary revascularization (OR 0.70; 95% CI 0.49–0.99). There appeared to be a lower risk of ischemic stroke of borderline statistical significance (OR 0.82; 95% CI 0.64–1.05). Heterogeneity ranged from low to moderate (I² = 26–72%). ACS acute coronary syndrome, CI confidence interval, MACE major adverse cardiovascular events, MI myocardial infarction, OR odds ratio

Effects of Colchicine on Mortality

A total of 13 studies evaluated all-cause mortality, with three reporting no deaths in either group (Fig. 4). The pooled analysis found no statistically significant differences in all-cause mortality between the colchicine and control groups (OR 0.98; 95% CI 0.82–1.19). Twelve studies reported cardiovascular mortality, with three noting no cardiovascular deaths in either group. Although a downward trend in cardiovascular mortality was observed with colchicine, the difference was not statistically significant (OR 0.92; 95% CI 0.73–1.15). Similarly, there were no statistically significant differences in non-cardiovascular mortality (OR 1.09; 95% CI 0.77–1.55), as reported in 11 studies. Among these, four reported zero non-cardiovascular deaths in either group. Although heterogeneity for all-cause or cardiovascular mortality was low, there was considerable heterogeneity for non-cardiovascular mortality (I² = 54%). These results are in keeping with previous concerns about colchicine in terms of non-cardiovascular mortality in some of these studies. These findings indicated that colchicine use did not significantly influence overall or non-cardiovascular mortality.

Fig. 4.

Effects of colchicine on mortality. Pooled analysis found no statistically significant differences in all-cause mortality between the colchicine and control groups (OR 0.98; 95% CI 0.82–1.19) or between cardiovascular mortality (OR 0.92; 95% CI 0.73–1.15) and non-cardiovascular mortality (OR 1.09; 95% CI 0.77–1.55). CI confidence interval, OR odds ratio.

Reduction of CRP with Colchicine

Five studies reported the effect of colchicine on hsCRP/CRP levels (Fig. 5). Pooled analyses revealed no significant mean differences in hsCRP/CRP levels between colchicine and control groups (pooled mean difference − 0.39 mg/L; 95% CI − 1.11 to 0.34). However, colchicine resulted in significantly lower post-treatment hsCRP/CRP levels (mean difference − 0.40 mg/L; 95% CI − 0.55 to − 0.25). Overall, there was a significant prospective reduction in hsCRP/CRP levels in patients receiving colchicine (mean difference − 0.27 mg/L; 95% CI − 0.53 to − 0.02). The results were robust among studies, with no heterogeneity observed (I² = 0%).

Fig. 5.

Effects of colchicine on hsCRP/CRP level. There were no significant differences in hsCRP/CRP levels between colchicine and control groups before treatment (MD − 0.39 mg/L; 95% CI − 1.11–0.34). Colchicine significantly reduced post-treatment hsCRP/CRP levels (MD − 0.40 mg/L; 95% CI − 0.55 to − 0.25) and resulted in prospective reductions in hsCRP/CRP levels (MD − 0.27 mg/L; 95% CI − 0.53 to − 0.02). CI confidence interval, CRP C-reactive protein, hsCRP highly sensitive CRP, MD mean difference, SD standard deviation.

Subgroup Analyses

Prespecified subgroup analyses further assessed the effect of colchicine across different primary disease subpopulations (Fig. 5 in the ESM). In total, 13 studies were included in the subgroup analysis according to the primary diseases. We excluded the COOL trial [31] because it included multiple populations. Results from the COLCHICINE-PCI trial [26] were analyzed for both the ACS and the MI subpopulations. In patients with stroke/TIA, colchicine treatment resulted in a trend of lower MACE, although this did not reach statistical significance (OR 0.90; 95% CI 0.77–1.05). Similarly lower risks of MACE with borderline statistical significance were seen in the subpopulations of MI (OR 0.89; 95% CI 0.79–1.02) and CAD (OR 0.47; 95% CI 0.21–1.08). In patients with ACS, there was a significant reduction of MACE (OR 0.39; 95% CI 0.25–0.63) with no heterogeneity (I² = 0%). These results indicate that colchicine may have stronger benefits for the prevention of MACE in patients with ACS; however, such discrepancies could be due to reduced sample sizes and study power in subgroup analyses.

Adverse Effects of Colchicine

Colchicine use was associated with several adverse effects. The most frequent of these were GI events such as nausea, diarrhea, and loose stools, as evidenced by a pooled OR of 1.95 (95% CI 1.29–2.94) derived from 10 studies (Fig. 6 in the ESM). However, there was substantial heterogeneity among studies reporting these events (I² = 92%). Nevertheless, serious GI adverse effects, including GI hemorrhage and hospitalization due to GI complications, were not significantly different between colchicine and control groups (OR 1.12; 95% CI 0.89–1.41), with no heterogeneity (I² = 0%).

For non-GI adverse effects, no significant differences were observed between the colchicine and control groups (Fig. 7 in the ESM). Infections were reported in seven studies, with a pooled OR of 1.03 (95% CI 0.85–1.24). Deranged liver or renal function, assessed in six studies, had a pooled OR of 1.24 (95% CI 0.83–1.83). Myalgia, reported by five studies, had an OR of 0.59 (95% CI 0.13–2.73). Neuropathy and myelosuppression, reported by four studies, had ORs of 1.21 (95% CI 0.42–3.52) and 0.52 (95% CI 0.05–6.05), respectively. Heterogeneity for these non-GI adverse effects ranged from low to moderate (I² = 22–58%), indicating overall consistency between studies. These findings indicated that colchicine use significantly increased the frequency of GI side effects, but it did not elevate the risk of serious complications.

Meta-Regression and Heterogeneity

We conducted meta-regression analyses to explore the sources of heterogeneity in the primary outcomes in all CVD populations and the incidence of GI adverse effects because of significant heterogeneities.

In the univariate meta-regression analysis for MACE recurrence (Table 1 in the ESM), later study year (OR 1.08; 95% CI 1.04–1.13, p = 0.0002) and larger sample size (OR 1.00008; 95% CI 1.00004–1.0001, p = 0.0003) were significantly associated with a slightly higher risk of MACE recurrence. To adjust for potential confounding and collinearity, we constructed a multivariable meta-regression model, including covariates that were statistically significant in univariate analysis, clinically important, and known to influence baseline CVD risk. In the multivariable model, none of the included factors remained statistically significant, suggesting that the observed univariate associations may be partially explained by underlying correlations among variables.

In meta-regression analysis for GI events (Table 2 in the ESM), study design significantly influenced the results: open-labelled RCTs were independently associated with a higher risk of detecting GI adverse effects (OR 3.95; 95% CI 2.06–7.60). Other factors, including sample size, duration and dosage of colchicine, sex, age, primary disease population, prevalence of hypertension and diabetes, and smoking status, did not significantly affect the heterogeneity observed in GI adverse effects. Further stratified analyses corroborated these findings (Fig. 8 in the ESM), with open-labelled RCTs reporting a higher incidence of GI adverse effects (OR 7.34; 95% CI 1.76–30.55) than double-blind RCTs (OR 1.41; 95% CI 1.09–1.83) (p-value for between-groups differences = 0.03).

Sensitivity Analysis

We conducted leave-one-out analyses to assess the statistical robustness of the primary findings. For patients with acute atherothrombotic CVDs (Fig. 9 in the ESM), the pooled OR remained statistically significant when any single study was eliminated, ranging between 0.74 and 0.87, indicating that no single study significantly dominated the overall result. Similarly, for patients in all CVD populations (Fig. 10 in the ESM), the pooled OR consistently ranged between 0.67 and 0.78, with no single study dominating the primary findings. We also performed sensitivity analyses by excluding studies with “some concerns” in the ROB2 assessment (LoDoCo-MI, COLCOT, and LoDoCo) (Fig. 11 in the ESM). The pooled ORs remained statistically significant and comparable in magnitude, supporting the robustness of our results. To assess publication bias, we used a funnel plot (Fig. 12 in the ESM), which showed no significant asymmetry. Egger test results (p = 0.06) indicated no significant evidence of publication bias.

Discussion

To our knowledge, this is the largest meta-analysis of RCTs to date to investigate the secondary protective effects of colchicine on MACE and related clinical safety, covering worldwide ethnicities. This is also the first study to consider the cumulative exposure of colchicine in terms of a dosage–time composite. We considered acute atherothrombotic CVD as a pathophysiological paradigm entailing coronary and cerebrovascular disease phenotypes, and all CVD in general. Findings of this meta-analysis indicated that colchicine significantly reduced MACE in both acute atherothrombotic and all CVD across multiple ethnicities, with a threshold protective effect that clinically corresponds to treatment exposure of 0.5 mg daily for ≥ 6 months. Importantly, there was no signal of increased all-cause mortality.

Based largely on the clinical efficacy of colchicine for secondary CVD prevention in patients with MI (COLCOT) [28] and stable CAD (LoDoCo and LoDoCo2) [24, 30], colchicine was incorporated in the European Society of Cardiology clinical guidelines in 2021 [32]. Under a class IIb recommendation, use of colchicine may be considered for secondary CVD prevention at a daily dose of 0.5 mg, particularly when CVD events are recurrent despite maximally tolerated conventional treatments or when other risk factors remain uncontrolled. The clinical evidence was assessed as level A. In 2023, the American Heart Association/American College of Cardiology clinical guidelines further recommended colchicine to be added to guideline-directed therapy in patients with chronic coronary disease for the secondary prevention of cardiovascular events (class IIb recommendation, level B evidence) [33]. Similar national guidelines have been implemented in Canada and South America [34]. Nevertheless, after the adoption of these clinical guidelines, several large RCT reported opposing (negative) results. These included patients with ischemic stroke (CONVINCE and CHANCE-3) [9, 10] and those with MI (CLEAR) [8]. Although these studies yielded trends toward a potential benefit of colchicine, none were statistically significant. For CLEAR, the COVID-19 pandemic had a notable impact on patient recruitment, necessitating a modification of the patient recruitment protocol. ST-segment elevation MI affected 95% of the patient sample. Although CLEAR concluded that the findings “seemed to be similar” during different phases relating to the COVID-19 pandemic, those enrolled before the pandemic appeared to give potential signals of protection with colchicine against the primary endpoint of cardiovascular death and events (hazard ratio 0.78; 95% CI 0.60–1.02), although this was statistically insignificant. Incorporating these new clinical data, our meta-analysis revealed that the overall body of RCT evidence to date affirms a protective effect of colchicine in secondary CVD prevention, which supports current guideline recommendations. The results were consistent between patients with acute atherothrombotic CVD and those with all CVD in general. The pooled estimate of risk reduction of MACE was 30% in our study, which also closely aligned with previous estimates of several landmark studies [24, 28, 34].

Interestingly, our analysis suggested that the cumulative dose of colchicine received over time may have important bearings on its ultimate clinical effectiveness as a secondary preventive agent. Here, ≥ 90 mg-days was defined a priori as the cut-off for comparison, as this was the median composite exposure of the overall study population. Conjunctively, it also represented a clinically meaningful guideline-directed daily dose of 0.5 mg colchicine [32, 34] for ≥ 6 months [35]. The cardiovascular protective effects of colchicine appear to accumulate progressively over time, reflecting a time-dependent pattern similar to that observed with statin therapy. This was demonstrated in a meta-analysis showing that statin therapy did not significantly reduce short-term events but led to benefits over longer follow-up [36]. The varying cumulative dose of colchicine used in previous studies may partially explain some of their conflicting results, especially in those with a relatively short treatment duration [10].

Colchicine significantly reduced individual components of MACE, including ACS, MI, unstable angina, and coronary revascularization procedures. These are generally in line with prior study findings [6, 11, 12]. Previous meta-analyses have reported varied results on specific disease phenotypes. A reduced risk of stroke recurrence was consistently observed in studies involving patients with ACS and CAD [6, 11–13], whereas a recent meta-analysis in patients with MI reported a non-significant result [14]. Our subgroup analyses suggest that the secondary protective effects of colchicine are conferred preferentially to patients with prior ACS, although this observation remained exploratory. For instance, a recent meta-analysis found that colchicine reduced ischemic stroke and MACE in patients with prior stroke or coronary diseases [37]. Nevertheless, it did not include two of the largest RCTs (CLEAR and CHANCE-3), both of which had yielded negative results for colchicine for secondary CVD prevention [8, 10]. Another meta-analysis reported a reduction in MACE with colchicine use, but it also excluded CHANCE-3 [38]. Furthermore, the sample size of our current meta-analysis is more than double that of earlier reports. Overall, it is reassuring that our findings are in agreement after taking into consideration these two negative trials. Our study is therefore both novel and timely. First, it reconciles major findings from the rapidly emerging body of RCT evidence on the secondary protective effects of colchicine in CVD. Second, it provides new insights into the cumulative dose effect of colchicine.

This meta-analysis of RCTs confirms that colchicine reduces serum CRP, a marker of systemic inflammation in patients with CVD. Colchicine modulates the upstream IL-1β/NLRP3 inflammasome-driven pathways [35], and this could underlie mechanisms for the secondary protection against atherosclerosis. Recent studies also showed further supporting clinical data for the role of colchicine on slowing atherosclerosis, especially in the setting of preponderance to inflammatory stressors. Although our observed absolute CRP reduction of − 0.27 mg/L appeared modest, other trials have shown that lowering CRP to <2 mg/L is associated with significant cardiovascular benefits [4, 5]. In the JUPITER trial [4], rosuvastatin reduced CRP by 37% and lowered cardiovascular event rates, even in patients without hyperlipidemia; in the CANTOS trial, canakinumab, an IL-1β monoclonal antibody, reduced CRP to a similar extent and significantly reduced recurrent cardiovascular events in patients with elevated CRP after MI, particularly among those with CRP levels < 2 mg/L [5]. Importantly, both trials also demonstrated cardiovascular risk reduction across the overall treatment groups, even among patients who did not reach the 2 mg/L threshold [4, 5]. This suggested that even modest CRP reductions may result in cardiovascular protection, with greater benefit with CRP levels <2 mg/L. However, rosuvastatin is primarily indicated for lipid management, and most patients with established CVD are already on statins, limiting its utility for further CVD prevention. The use of canakinumab was associated with increased rates of fatal infections [5]. The drug cost involved was high. Colchicine has broader pharmacological properties than rosuvastatin and canakinumab, is clinically widely available, and has a very low drug cost. Our meta-analysis found that colchicine does not increase the risk of infection (OR 0.97; 95% CI 0.84–1.12), in keeping with previous studies [6, 11, 12].

Furthermore, previous studies noted a concerning increased risk of non-cardiovascular death in patients with chronic coronary disease treated with colchicine [24]. Reassuringly, this meta-analysis showed no significant changes in all-cause or non-cardiovascular mortality. Our meta-analysis benefited from an increased patient sample size, particularly by incorporating several recent large RCTs (CLEAR, CONVINCE, CHANCE-3) [8–10]. This may have reduced the probability of a spurious association. On the other hand, GI adverse effects with colchicine have been well described [11, 12, 14]. These are typically limiting. Our study found no significant increase in the risk of serious GI adverse effects. Clinical data suggested that the tolerability of colchicine may be improved through dose titration or co-administration of antimotility agents [39]. Importantly, those with pre-existing GI conditions, chronic liver disease, or renal impairment, who are more susceptible to colchicine toxicity, may not be suitable for colchicine use [34]. Patient-centered, individualized assessments for judicious use of colchicine, including exclusion of those with contraindications, remains fundamental [34]. Interestingly, GI adverse effects were more commonly associated with colchicine when the RCTs were open labelled. The lack of blinding may amplify the perception and reporting of subjective outcomes such as GI symptoms [40]. This highlights the importance of a double-blinding strategy in future RCTs using colchicine. Overall, our findings substantiate the general safety of colchicine as a therapeutic option in patients with CVD.

Study Limitations

This study has several important limitations that should be acknowledged. First, this was a meta-analysis of study-level data. Individual data analysis was not available. Furthermore, the definitions of MACE varied across the included studies, which may affect the ultimate comparability of results. Additionally, there were moderate heterogeneities in the primary outcomes. Although we conducted meta-regression to explore the sources of heterogeneity, these analyses did not explain all the variability. Including both open-labelled and double-blind RCTs also introduced potential biases, with outcomes from open-labelled studies being influenced by the absence of blinding. Although our search strategy included studies in multiple languages, all eligible studies identified were published in English, which may have introduced a risk of language bias. Moreover, recently published studies focusing on stroke and TIA populations suggested that this research area was relatively new, with limited data available to fully reflect the outcomes of colchicine in this population. Lastly, some included studies had a relatively short follow-up period, which might not adequately reflect long-term outcomes.

Conclusions

In this meta-analysis of RCTs including 31,397 patients, colchicine significantly reduced MACE in both acute atherothrombotic and all CVD across multiple ethnicities, with a threshold protective effect that clinically corresponds to treatment of 0.5 mg daily for ≥ 6 months. Importantly, there was no signal of increased all-cause mortality.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Funding

This study was supported by Li Shu Pui Medical Foundation Fellowship 2022; Seed Funding for Basic Research (104006735) and Start-Up Funding, The University of Hong Kong; and Hong Kong Research Grants Council General Research Fund (GRF 17110122 and 17108523). YHC was supported by the Hong Kong Research Grants Council Clinical Research Fellowship 2024/25, and Doris Zimmern HKU-Cambridge Hughes Hall Fellowship 2024/25.

Conflicts of Interest:

Li HY, Cheriyan J, Chan TK, Yiu KH, Tse HF, Wilkinson IB, and Chan YH have no potential conflicts of interest that might be relevant to the contents of this manuscript.

Author Contributions

Li HY contributed to the research methodology, collected the data, performed the analyses, and wrote the manuscript. Cheriyan J and Yiu KH contributed clinical cardiology/pharmacology expertise, contributed to the research methodology, and co-authored the manuscript. Chan TK collected and analyzed the data. Tse HF and Wilkinson IB supervised the study direction, implementation and analyses, wrote and critically revised the manuscript, and are co-senior authors. Chan YH originated the study hypothesis, obtained research funding, designed and implemented the study, performed and supervised the analyses, wrote and revised the manuscript, and was the overall guarantor of the study. All authors wrote, revised, and approved the manuscript.

Data Availability Statement

The data used are from previously published studies. Derived data in this study are not publicly available but are available upon reasonable request to the corresponding author.

Ethics Approval

Not applicable.

Code Availability

Not applicable.

Consent to Participate

Not applicable.

Consent for Publication

Not applicable.